Money and Capital in an OLG Model D. Andolfatto June 2011
Environment Time is discrete and the horizon is infinite ( =1 2 ) At the beginning of time, there is an initial old population that lives (participates) for one period only At the beginning of every time period, a new generation of young agents is born (enter the economy) Let denote population of young born at date ; assume +1 =
Young agents live (participate) for two periods; they become old and then die (exit the economy) Anagentbornatdate =0 1 has preferences = +1 Young agents have an endowment All agents have access to an investment technology, where units of investment at date delivers ( ) at date +1 Capital depreciates fully after it is used in production Assume that is strictly increasing, concave, and satisfies the Inada conditions
Competitive equilibrium Let +1 denote the real (gross) rate of interest on risk-free debt No-arbitrage-condition implies 0 ( )= +1 The young save all their income, so = and +1 = ( ) for all 1 (the initial old consume zero) This implies an equilibrium real rate of interest = 0 ( ) The CE is not Pareto efficient if (dynamic inefficiency)
Gift-giving equilibria Suppose that the young are asked to make a gift of output [0 ] to the old In return, the young expect to receive units of output when they are old Payoff is = ( )+ Any 0 such that ( )+ ( ) can be supported as a Nash equilibrium with tit-for-tat strategies
Note: 0 iff 0 ( ) (Re: dynamic inefficiency); otherwise, only equilibrium is autarky Note: maximum payoff for representative young agent is achieved by setting to maximize = ( )+ so that 0 ( )= (1) If 0 ( ) then all such gift-giving equilibria constitute Pareto improvements over competitive equilibrium allocation Note: initial old consume 0
Supporting gift-giving equilibria with government debt In what follows, I make no distinction between money and bonds; i.e., a bond is like interest-bearing money moneyislikeazero-interestbond Let denote the nominal stock of money at date One dollar of money held at date constitutes a risk-free claim for +1 dollars of money in period +1 So +1 is the (gross) nominal interest rate on government money/debt
Assume that the initial stock of money 0 is distributed evenly across the initial old (so 0 0 dollars per initial old agent) note: the old (of any generation) are willing to dispose of their money holdings at any price The money supply grows according to +1 = +1 There is a government budget constraint ( 1) 1 =( 1 )+ interest on debt is financed by new money creation and/or taxes Lump-sum tax (transfer, if negative) is applied to the old
Rearrange GBC... Rewrite GBC as 1 = + From money growth equation, 1 = ; plug into GBC above to get " # = 1 (2) Note special cases: ( ) {(1 ) ( 1) (1 1)} Note: usually, the restriction 1 is imposed (idea is that currency in circulation cannot be taxed)
Individual budget constraints Let denote the price level at date Budget constraints for representative young agent = + +1 +1 = +1 ( )+ +1 +1 Define (real money balances) and +1 +1 ( +1 ) then rewrite above constraints as +1 = ( )+( +1 Π +1 ) +1 (3) where Π +1 +1 is the gross rate of inflation
Demand for money The demand for real money balances = ( +1 Π +1 ) satisfies (at an interior) 0 ( )= +1 Π +1 (4) This also implies a demand for capital investment, = Money/bond demand is increasing in and decreasing in Π
General equilibrium Money trades for goods on competitive spot market in each period; marketclearing condition implies (for all 1) = (5) Assume =1 = = for all 1 and define 1 Combining previous two restrictions... +1 = 0 ( ) ( ) which is a first-order difference equation in { } =1
Steady-state equilibria A steady-state equilibrium has the property = +1 = for all There are two steady-states, one monetary ( 0) and one nonmonetary ( =0) Monetary steady-state 0 satisfies 1= 0 ( ) Note: the monetary equilibrium is an asset bubble (fiat money is priced above its fundamental value)
Nonstationary price-level dynamics with constant fundamentals 0 ( ) = 0 ( ) 00 ( ) 0 00 ( ) = 2 00 ( )+ 000 ( ) Let s assume 000 0 so that 00 0 Equilibrium sequence { } must be bounded; i.e., 0 Then there exists a continnum of nonstationary monetary equilibria indexed by initial condition 1 (0 ) where & 0 (inflation)
Exercise 1 Does there exist an interest rate policy rule, say, in the form +1 = (Π ) that will ensure long-run stability of the price-level? If so, describe the properties of this policy (or any other policy that may do the trick).
Stationary equilibria Let = 1 and = with = 1 Then from market-clearing condition (5) +1 +1 = " # +1 +1 Π +1 = +1 In a stationary equilibrium, = +1 so that Π +1 = for all So conditional on a policy ( ) with associated lump-sum tax/transfer to balance budget, the stationary monetary equilibrium is characterized as follows...
Equilibrium inflation rate is ˆΠ = From condition (4) we have equilibrium quantity of real money balances 0 ( ˆ ) = (6) Initial price-level is given by 1 = 1ˆ From GBC (2) we have ˆ =[( ) 1] ˆ From BC (3) we have ˆ = ( ˆ )+ ˆ Note ˆ = for any policy with = (compare conditions 1 and 6)
Money neutrality Moneyissaidtobeneutral if a one-time change in has no real effects (i.e., if only nominal values change) Money is said to be superneutral if a one-time change in has no real effects Money is clearly not superneutral in this model But money may or may not be neutral Depends on how new money is injected into the economy
Note: this is a heterogeneous agent economy (the young and old have different consumption propensities) A lump-sum injection of cash to the old turns out to be neutral the old compete among themselves and bid up the price of goods in proportion to their new money A lump-sum injection of cash to the young is neutral in the long-run, but notintheshort-run new money puts upward pressure on the price-level and diminishes the purchasing power of cash held by old wealth transfer from old to young (and a temporary investment boom)
Exercise 2 Work out the economic consequences of a one-time increase in the money supply at date 1, where the new money is injected as a lump-sum transfer to the initial young.
Inflation as a tax and the limits to seigniorage Let =1and = 1 so that =[1 1] Define seigniorage so that =[1 1 ] Convert into real terms and divide by 1 = =[1 1 ] Ã! So is real seigniorage revenue per old agent
Looks like is an increasing function of ; but is this necessarily the case? We know that condition (6) must hold; i.e., 0 ( ˆ ) = Since ˆ = in a stationary equilibrium, we have ˆ ( ) =[1 1 ] ˆ ( ) Note that the real tax base ˆ ( ) is a decreasing function of Consequently, ˆ ( ) exhibits a Laffer curve property
An increase in the money supply that has no price-level effect An increase in the money supply need not have any effect (including nominal effects) if the increase is temporary Example: a lump-sum injection of cash at some arbitrary date (to the young at date ) that is to be reversed at date +1(via a lump-sum taxontheoldatdate +1) Such a policy has no effect on the lifetime wealth of young agents; they will simply save to pay back an anticipated tax bill next period The demand for money rises one-for-one with the supply of money (Ricardian equivalence holds in this case)
Exercise 3 Repeat the experiment above except assume now that the lumpsum money transfer accrues to the old, instead of the young. Explain why Ricardian equivalence does or does not hold in this case.
Monetary policy as asset purchases In reality, central banks expand/contract the supply of base money by way of purchases/sales of assets Traditionally, the assets involved are typically short-term government debt instruments (e.g., U.S. Treasuries) Occasionally, other asset classes are involved; e.g., long-term government debt and mortgage-backed securities Collateralized lending arrangements (e.g., the discount window), sale and repurchase agreements (repo), are closely related policies
Purchases of capital Think of the Fed s massive MBS purchase program what does our theory say? Let s begin in a steady-state Government then steps in, prints new money and uses it to purchase output (in the form of capital, ) Money supply will be permanently higher, unless the government plans to reverse the money injection
One way to reverse is to re-sell the capital (or its output, in this case) for money in the future Such a policy is likely to be completely neutral here (I have not checked) Upshot: the inflationary consequences of an increase in the supply of base money depends critically on whether the increase is (or is expected to be) reversed in the future Note: the Fed has repeatedly suggested that it has an exit strategy in place
Broad money aggregates and the money multiplier Most of an economy s money supply consists of private debt In reality, debt in the form of demandable liabilities (demand deposit liabilities) Broad money measure M1 defined roughly as cash plus demand deposits We can identify a similar (though not identical) object in our model In our model, think of as work effort devoted to producing new capital goods
Imagine that a firm borrows banknotes from a bank to pay for Banknotes constitute a claim against future output; worker deposits banknotes back into bank Young worker uses banknotes to purchase future output from firm, firm uses sales to pay back loan to bank Nominal value of private money is The aggregate money supply is 1 = + From the market-clearing condition = ( )
Combining the previous two expressions... 1 = " 1+ # The money mulitiplier is defined as 1 ; i.e., the object in square brackets Anything that causes to expand (or to contract) will cause an expansion in the broad money supply even if remains constant (banks are expanding their balance sheets) Conversely, any shock that induces a large increase in the demand for government money (at the expense of private money) will cause a decline in the supply of broad money (banks are contracting their balance sheets)
U.S. money multiplier plunges in the great recession...
Excess money demand? Many economists/commentators explain recessions as the outcome of some shock that causes an excess demand for money (and/or government securities) People get scared (irrationally?) and collectively flock into relatively safe assets (dumping risky assets) Now, the flight into money is particularly troubling because, if nominal prices are sticky, there is now not enough money to facilitate a normal level of transactions, so output must decline (an excess supply of output is the other side of an excess demand for money)
Our model offers a slightly different interpretation that does not rely on sticky prices Modify the investment technology as follows [ +1 ( ) ] where is a parameter that indexes the productivity of capital investment, and where is information (or psychology) that influences conditional forecasts Assume that [ +1 ] is increasing in and that follows an exogenous stochastic process
Our theory implies [ +1 ] 0 ( ) = [ +1 ] Now, invoke the market-clearing condition = ; so that Ã! [ +1 ] 0 ( ) = [ +1 ] Iconjecturethat = ( ) To simplify, assume that is an i.i.d. process (might be interesting to generalize later); consequently h ( +1 ) i is a constant and [ +1 ] 0 ( ( )) ( )= Ã!
Note that 0 ( ) is increasing in Consequently, ( ) must be decreasing in The implication is that bad news (or psychological depression) will lead to an increase in the demand for money (asset substitution away from private money) That is, ( ) falls money multiplier falls Moreover, this type of shock is deflationary (the price level jumps down)
U.S. price level (CPI)
Policy responses If the flight to quality is deemed excessive (driven perhaps by animal spirits), then an appropriate policy response is to lower the interest rate Re: is an increasing function of in this model (and lower stimulates investment) If there is a constraint 1 then there is only so far this policy can go Some economists (e.g., Paul Krugman) argue that increasing is desirable if =1
Unconventional policy responses Conventional monetary policy organized around a Taylor rule = + (Π Π)+ ( ) But constrained by 1 So try unconventional measures, like quantitative easing (QE) Not as unconventional as people make out Basic idea is to print money to finance large-scale purchases of other asset classes (mortgage-backed securities and longer-dated Treasuries)
The impact of these QE policies is hotly debated There is some evidence to suggest they increased inflation expectations (at a time when deflation was a worry) QE2 formally announced in early November 2010 (hinted earlier in summer at Jackson Hole)